Job history server in Hadoop 2.7.1 is not working

Question

I explored everywhere but could not get the solution. When I do run jps command it does not show job history server running.

What are configuration settings for history server and how to start this?

EDIT 1:

When history server is stopped, all previously recorded information of jobs is deleted. How to retain all those information after cluster and/or history server is restarted?

Answer 1

JobHistoryServer process has to be started separately.

To Start:

$HADOOP_HOME/sbin/mr-jobhistory-daemon.sh start historyserver

To Stop:

$HADOOP_HOME/sbin/mr-jobhistory-daemon.sh stop historyserver

If the Cluster setup is Pseudo Distributed, no additional properties are required.

In a multinode cluster setup, Add these JobHistoryServer properties in mapred-site.xml:

<property>
   <name>mapreduce.jobhistory.address</name>
   <value>hostname:10020</value> <!-- hostname of machine  where jobhistory service is started -->
</property>
<property>
   <name>mapreduce.jobhistory.webapp.address</name>
   <value>hostname:19888</value> 
</property>